D.S. Anisimov, M.A. Ryazanov, A.I. Chapoval
Approaches for Large Volume Data Analysis Obtained Utilizing Peptide Microarrays
This paper discusses strategies for high density peptide microarray data analysis. Main steps of the analysis include preprocessing, normalization (the reduction of hardware measurement errors), data volume reduction (the selection of statistically significant variables which describe the original data the best) and data classification (class determination among multiple samples). The study was designed for different algorithms optimization and testing to analyze large volume data. The main objectives of the study were improving existing algorithms and enhancing data sustainability for these algorithms utilization to analyze data obtained with peptide microarray. The work was performed on a relatively small sample set (25 donors, 15 — healthy and 10 — with breast cancer). The result of this study is a developed technology for peptide microarray data analysis, which may be used to evaluate a large number of samples and a greater number of classes.
Key words: peptide microchip, processing of multidimensional data, ordinary least squares, support vector machines, naive Bayes classifier, k-nearest neighbors
Full text at PDF
, 919Kb. Language: Russian.